#include <linux/swap.h>
#include <linux/writeback.h>
#include <linux/task_io_accounting_ops.h>
-#include <linux/interrupt.h>
-#include <linux/cpu.h>
#include <linux/blktrace_api.h>
#include <linux/fault-inject.h>
*/
static struct workqueue_struct *kblockd_workqueue;
-static DEFINE_PER_CPU(struct list_head, blk_cpu_done);
-
static void drive_stat_acct(struct request *rq, int new_io)
{
struct hd_struct *part;
if (!blk_fs_request(rq) || !rq->rq_disk)
return;
- cpu = disk_stat_lock();
+ cpu = part_stat_lock();
part = disk_map_sector_rcu(rq->rq_disk, rq->sector);
if (!new_io)
- all_stat_inc(cpu, rq->rq_disk, part, merges[rw], rq->sector);
+ part_stat_inc(cpu, part, merges[rw]);
else {
- disk_round_stats(cpu, rq->rq_disk);
- rq->rq_disk->in_flight++;
- if (part) {
- part_round_stats(cpu, part);
- part->in_flight++;
- }
+ part_round_stats(cpu, part);
+ part_inc_in_flight(part);
}
- disk_stat_unlock();
+ part_stat_unlock();
}
void blk_queue_congestion_threshold(struct request_queue *q)
memset(rq, 0, sizeof(*rq));
INIT_LIST_HEAD(&rq->queuelist);
- INIT_LIST_HEAD(&rq->donelist);
+ rq->cpu = -1;
rq->q = q;
rq->sector = rq->hard_sector = (sector_t) -1;
INIT_HLIST_NODE(&rq->hash);
blk_add_trace_pdu_int(q, BLK_TA_UNPLUG_TIMER, NULL,
q->rq.count[READ] + q->rq.count[WRITE]);
- kblockd_schedule_work(&q->unplug_work);
+ kblockd_schedule_work(q, &q->unplug_work);
}
void blk_unplug(struct request_queue *q)
}
EXPORT_SYMBOL(blk_unplug);
+static void blk_invoke_request_fn(struct request_queue *q)
+{
+ /*
+ * one level of recursion is ok and is much faster than kicking
+ * the unplug handling
+ */
+ if (!queue_flag_test_and_set(QUEUE_FLAG_REENTER, q)) {
+ q->request_fn(q);
+ queue_flag_clear(QUEUE_FLAG_REENTER, q);
+ } else {
+ queue_flag_set(QUEUE_FLAG_PLUGGED, q);
+ kblockd_schedule_work(q, &q->unplug_work);
+ }
+}
+
/**
* blk_start_queue - restart a previously stopped queue
* @q: The &struct request_queue in question
WARN_ON(!irqs_disabled());
queue_flag_clear(QUEUE_FLAG_STOPPED, q);
-
- /*
- * one level of recursion is ok and is much faster than kicking
- * the unplug handling
- */
- if (!queue_flag_test_and_set(QUEUE_FLAG_REENTER, q)) {
- q->request_fn(q);
- queue_flag_clear(QUEUE_FLAG_REENTER, q);
- } else {
- blk_plug_device(q);
- kblockd_schedule_work(&q->unplug_work);
- }
+ blk_invoke_request_fn(q);
}
EXPORT_SYMBOL(blk_start_queue);
* Only recurse once to avoid overrunning the stack, let the unplug
* handling reinvoke the handler shortly if we already got there.
*/
- if (!elv_queue_empty(q)) {
- if (!queue_flag_test_and_set(QUEUE_FLAG_REENTER, q)) {
- q->request_fn(q);
- queue_flag_clear(QUEUE_FLAG_REENTER, q);
- } else {
- blk_plug_device(q);
- kblockd_schedule_work(&q->unplug_work);
- }
- }
+ if (!elv_queue_empty(q))
+ blk_invoke_request_fn(q);
}
EXPORT_SYMBOL(__blk_run_queue);
__elv_add_request(q, req, ELEVATOR_INSERT_SORT, 0);
}
-/*
- * disk_round_stats() - Round off the performance stats on a struct
+static void part_round_stats_single(int cpu, struct hd_struct *part,
+ unsigned long now)
+{
+ if (now == part->stamp)
+ return;
+
+ if (part->in_flight) {
+ __part_stat_add(cpu, part, time_in_queue,
+ part->in_flight * (now - part->stamp));
+ __part_stat_add(cpu, part, io_ticks, (now - part->stamp));
+ }
+ part->stamp = now;
+}
+
+/**
+ * part_round_stats() - Round off the performance stats on a struct
* disk_stats.
*
* The average IO queue length and utilisation statistics are maintained
* /proc/diskstats. This accounts immediately for all queue usage up to
* the current jiffies and restarts the counters again.
*/
-void disk_round_stats(int cpu, struct gendisk *disk)
-{
- unsigned long now = jiffies;
-
- if (now == disk->stamp)
- return;
-
- if (disk->in_flight) {
- disk_stat_add(cpu, disk, time_in_queue,
- disk->in_flight * (now - disk->stamp));
- disk_stat_add(cpu, disk, io_ticks, (now - disk->stamp));
- }
- disk->stamp = now;
-}
-EXPORT_SYMBOL_GPL(disk_round_stats);
-
void part_round_stats(int cpu, struct hd_struct *part)
{
unsigned long now = jiffies;
- if (now == part->stamp)
- return;
-
- if (part->in_flight) {
- part_stat_add(cpu, part, time_in_queue,
- part->in_flight * (now - part->stamp));
- part_stat_add(cpu, part, io_ticks, (now - part->stamp));
- }
- part->stamp = now;
+ if (part->partno)
+ part_round_stats_single(cpu, &part_to_disk(part)->part0, now);
+ part_round_stats_single(cpu, part, now);
}
+EXPORT_SYMBOL_GPL(part_round_stats);
/*
* queue lock must be held
void init_request_from_bio(struct request *req, struct bio *bio)
{
+ req->cpu = bio->bi_comp_cpu;
req->cmd_type = REQ_TYPE_FS;
/*
req->biotail = bio;
req->nr_sectors = req->hard_nr_sectors += nr_sectors;
req->ioprio = ioprio_best(req->ioprio, prio);
+ if (!blk_rq_cpu_valid(req))
+ req->cpu = bio->bi_comp_cpu;
drive_stat_acct(req, 0);
if (!attempt_back_merge(q, req))
elv_merged_request(q, req, el_ret);
req->sector = req->hard_sector = bio->bi_sector;
req->nr_sectors = req->hard_nr_sectors += nr_sectors;
req->ioprio = ioprio_best(req->ioprio, prio);
+ if (!blk_rq_cpu_valid(req))
+ req->cpu = bio->bi_comp_cpu;
drive_stat_acct(req, 0);
if (!attempt_front_merge(q, req))
elv_merged_request(q, req, el_ret);
init_request_from_bio(req, bio);
spin_lock_irq(q->queue_lock);
+ if (test_bit(QUEUE_FLAG_SAME_COMP, &q->queue_flags) ||
+ bio_flagged(bio, BIO_CPU_AFFINE))
+ req->cpu = blk_cpu_to_group(smp_processor_id());
if (elv_queue_empty(q))
blk_plug_device(q);
add_request(q, req);
out:
if (sync)
__generic_unplug_device(q);
-
spin_unlock_irq(q->queue_lock);
return 0;
static int should_fail_request(struct bio *bio)
{
- if ((bio->bi_bdev->bd_disk->flags & GENHD_FL_FAIL) ||
- (bio->bi_bdev->bd_part && bio->bi_bdev->bd_part->make_it_fail))
+ struct hd_struct *part = bio->bi_bdev->bd_part;
+
+ if (part_to_disk(part)->part0.make_it_fail || part->make_it_fail)
return should_fail(&fail_make_request, bio->bi_size);
return 0;
struct hd_struct *part;
int cpu;
- cpu = disk_stat_lock();
+ cpu = part_stat_lock();
part = disk_map_sector_rcu(req->rq_disk, req->sector);
- all_stat_add(cpu, req->rq_disk, part, sectors[rw],
- nr_bytes >> 9, req->sector);
- disk_stat_unlock();
+ part_stat_add(cpu, part, sectors[rw], nr_bytes >> 9);
+ part_stat_unlock();
}
total_bytes = bio_nbytes = 0;
return 1;
}
-/*
- * splice the completion data to a local structure and hand off to
- * process_completion_queue() to complete the requests
- */
-static void blk_done_softirq(struct softirq_action *h)
-{
- struct list_head *cpu_list, local_list;
-
- local_irq_disable();
- cpu_list = &__get_cpu_var(blk_cpu_done);
- list_replace_init(cpu_list, &local_list);
- local_irq_enable();
-
- while (!list_empty(&local_list)) {
- struct request *rq;
-
- rq = list_entry(local_list.next, struct request, donelist);
- list_del_init(&rq->donelist);
- rq->q->softirq_done_fn(rq);
- }
-}
-
-static int __cpuinit blk_cpu_notify(struct notifier_block *self,
- unsigned long action, void *hcpu)
-{
- /*
- * If a CPU goes away, splice its entries to the current CPU
- * and trigger a run of the softirq
- */
- if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
- int cpu = (unsigned long) hcpu;
-
- local_irq_disable();
- list_splice_init(&per_cpu(blk_cpu_done, cpu),
- &__get_cpu_var(blk_cpu_done));
- raise_softirq_irqoff(BLOCK_SOFTIRQ);
- local_irq_enable();
- }
-
- return NOTIFY_OK;
-}
-
-
-static struct notifier_block blk_cpu_notifier __cpuinitdata = {
- .notifier_call = blk_cpu_notify,
-};
-
-/**
- * blk_complete_request - end I/O on a request
- * @req: the request being processed
- *
- * Description:
- * Ends all I/O on a request. It does not handle partial completions,
- * unless the driver actually implements this in its completion callback
- * through requeueing. The actual completion happens out-of-order,
- * through a softirq handler. The user must have registered a completion
- * callback through blk_queue_softirq_done().
- **/
-
-void blk_complete_request(struct request *req)
-{
- struct list_head *cpu_list;
- unsigned long flags;
-
- BUG_ON(!req->q->softirq_done_fn);
-
- local_irq_save(flags);
-
- cpu_list = &__get_cpu_var(blk_cpu_done);
- list_add_tail(&req->donelist, cpu_list);
- raise_softirq_irqoff(BLOCK_SOFTIRQ);
-
- local_irq_restore(flags);
-}
-EXPORT_SYMBOL(blk_complete_request);
-
/*
* queue lock must be held
*/
struct hd_struct *part;
int cpu;
- cpu = disk_stat_lock();
+ cpu = part_stat_lock();
part = disk_map_sector_rcu(disk, req->sector);
- all_stat_inc(cpu, disk, part, ios[rw], req->sector);
- all_stat_add(cpu, disk, part, ticks[rw], duration, req->sector);
- disk_round_stats(cpu, disk);
- disk->in_flight--;
- if (part) {
- part_round_stats(cpu, part);
- part->in_flight--;
- }
+ part_stat_inc(cpu, part, ios[rw]);
+ part_stat_add(cpu, part, ticks[rw], duration);
+ part_round_stats(cpu, part);
+ part_dec_in_flight(part);
- disk_stat_unlock();
+ part_stat_unlock();
}
if (req->end_io)
struct request_queue *q = rq->q;
unsigned long flags = 0UL;
- if (bio_has_data(rq->bio) || blk_discard_rq(rq)) {
+ if (rq->bio) {
if (__end_that_request_first(rq, error, nr_bytes))
return 1;
**/
int __blk_end_request(struct request *rq, int error, unsigned int nr_bytes)
{
- if ((bio_has_data(rq->bio) || blk_discard_rq(rq)) &&
- __end_that_request_first(rq, error, nr_bytes))
+ if (rq->bio && __end_that_request_first(rq, error, nr_bytes))
return 1;
add_disk_randomness(rq->rq_disk);
rq->rq_disk = bio->bi_bdev->bd_disk;
}
-int kblockd_schedule_work(struct work_struct *work)
+int kblockd_schedule_work(struct request_queue *q, struct work_struct *work)
{
return queue_work(kblockd_workqueue, work);
}
int __init blk_dev_init(void)
{
- int i;
-
kblockd_workqueue = create_workqueue("kblockd");
if (!kblockd_workqueue)
panic("Failed to create kblockd\n");
blk_requestq_cachep = kmem_cache_create("blkdev_queue",
sizeof(struct request_queue), 0, SLAB_PANIC, NULL);
- for_each_possible_cpu(i)
- INIT_LIST_HEAD(&per_cpu(blk_cpu_done, i));
-
- open_softirq(BLOCK_SOFTIRQ, blk_done_softirq);
- register_hotcpu_notifier(&blk_cpu_notifier);
-
return 0;
}